Men who have sex with men: estimating the size of at-risk populations in London primary care trusts

Abstract

The objective of this study was to estimate local at-risk populations of men who have sex with men (MSM) in London primary care trusts (PCTs) to inform the commissioning of targeted health interventions. Estimated population size and prevalence of diagnosed HIV in MSM in all of London were calculated using data from the British National Survey of Sexual Attitudes and Lifestyles (NATSAL), Greater London Authority population estimates and the annual survey of diagnosed MSM (Survey of Prevalent HIV Infections Diagnosed [SOPHID]). Estimated MSM population sizes at the PCT level were calculated using un-weighted and SOPHID-weighted methods and methods discussed. Four-fifths of MSM with diagnosed HIV infection in Greater London lived in inner London. Estimated population size of MSM 16–44 years in inner London was 66,000; estimated overall prevalence of diagnosed HIV infection among MSM was 9.5%. Our models show substantial variation at the PCT level between the two methods. Using the SOPHID-weighted method MSM account for up to 16% of the male population in some London PCTs, compared with as low as 3% in others. We provide a novel method of estimating at-risk MSM populations living in inner London PCTs indicating that proportions of MSM vary widely between PCTs. Significant proportions of MSM among the resident populations in several PCTs warrant inclusion of MSM health needs in core PCT prevention and service programming. In light of data source limitations further validation studies are needed.

Keywords

HIV prevention homosexuality epidemiology inequalities MSM at-risk population sexually transmitted infection health-care planning UK

BACKGROUND

The planning and commissioning of local sexual health services and prevention programmes requires an understanding of populations most affected by and at risk of HIV and other sexually transmitted infections (STIs). In the UK, there is paucity of good local information about population characteristics, such as sexuality, which are highly relevant for the planning of prevention and services. Almost three decades into the HIV epidemic, men who have sex with men (MSM) remain the group most affected by and at risk of HIV in the UK, with around half of the estimated 25,600 diagnosed HIV-infected MSM living within the capital.¹ New HIV diagnoses among MSM continue to increase and over four-fifths of these infections were probably acquired in the UK.

UK primary care trusts (PCTs) are geographically defined devolved health organizations that commission and provide most of the primary and secondary health care for local populations. London PCTs are congruent with borough boundaries. There is limited information on the MSM at-risk population size in the UK, as sexuality is currently not recorded in UK routine population level information systems. The British National Survey of Sexual Attitudes and Lifestyles 2000 (NATSAL II) provides the best estimates of the prevalence of MSM aged 16–44 years for England and Greater London. In 2000, an estimated 2.6% (95% confidence interval [CI] 2.2–3.1%) in England and 5.5% (95% CI 4.2–7.1%) in Greater London of men aged 16–44 years and reported same-sex contact in the preceding five years (the definition used for MSM).² However, within Greater London the proportion of MSM was almost twice as high in inner London (8.27% [95% CI 6.36–10.69%]; sample size: 744 unweighted) compared with outer London (3.72% [95% CI 2.23–6.14%]; sample size: 590 unweighted) (unpublished data from NATSAL II, personal communication, Dr Catherine Mercer). However, the sample size of the survey is insufficient to provide smaller geographical breakdowns of MSM population below inner London as a whole (e.g. at borough/PCT level). Furthermore, evidence from community-based surveys suggests that the distribution of MSM living in London is not homogeneous, with many gay men choosing to live in certain parts of the capital, and it is unlikely that the NATSAL II estimate for inner London overall (8.3%) is generalizable for all inner PCTs. These data, currently lacking in local epidemiological profiles and needs assessments, are critically important for the commissioning of targeted local sexual health interventions aimed at local at-risk MSM populations.

In this paper we set out to utilize existing data sources relating to HIV epidemiology, behavioural information and population estimates and explore their utility in determining the size of local MSM populations living within London.

METHODS

We used the geographical distribution of diagnosed HIV-positive MSM of the Health Protection Agency's (HPA) Survey of Prevalent HIV Infections Diagnosed (SOPHID) data (in addition to other sources) as a novel approach to produce weighted estimates of the number of MSM residing within inner London PCTs and compared these with MSM population estimates based on NATSAL II findings alone.

Data sources

Survey of Prevalent HIV Infections Diagnosed

SOPHID is a voluntary annual cross-sectional survey of all individuals with diagnosed HIV infection accessing a National Health Service (NHS) for HIV-related care in England, Wales and Northern Ireland.³ Aggregate data on men who acquired their HIV infection through sex between men and accessed care in 2008 within inner London were used in these analyses.

National Survey of Sexual Attitudes and Lifestyles 2000 (NATSAL II)

NATSAL is a well-recognized in-depth survey of patterns of sexual behaviour in Britain. Details of the survey's methodology have been published elsewhere.²

Greater London Authority population estimates

Mid-year estimates of the male population aged 16–44 years from the Greater London Authority (GLA) Round Ward Population Projections 2008 – (RLP) High⁴ were used as the denominator for calculating PCTs' rates. The GLA population estimates were based on the UK census and projections from the Office of National Statistics (ONS). It is the model preferred by London PCTs for service planning as it is designed specifically for London whereas the ONS model is considered best to produce population projections for the whole of the UK. inner and outer London boroughs/PCTs were classified according the UK ONS classification.

Analytical approaches

Analyses were restricted to men aged 16–44 years in line with available NATSAL II estimates.

MSM population estimates aged 16–44 years for inner London

The estimated overall MSM population size for inner London was calculated by multiplying the estimated inner London MSM proportion (NATSAL II) by the 2008 inner London male population.

Overall prevalence of diagnosed HIV in MSM aged 16–44 years across inner London

The overall prevalence of diagnosed HIV infection among MSM aged 16–44 years in inner London was estimated by dividing the number of diagnosed HIV-infected MSM 16–44 years (SOPHID data) by the estimated MSM population size (aged 16–44 years).

MSM population estimates aged 16–44 years for individual inner London PCTs

Two methods were used to estimate the population size of MSM residing within each inner London PCT.

NATSAL II only Based Estimates (NoBE): Assuming a constant proportion of MSM across all inner London PCTs, the estimated proportion of MSM in inner London derived from NATSAL II (8.27% [6.36–10.69%]) was multiplied by the male population (aged 16–44 years) in each PCT (GLA);

SOPHID Weighted Estimates (SWE): Assuming a constant overall prevalence of diagnosed HIV infection among MSM across inner London as calculated above, the number of MSM reported to SOPHID in each of the inner London PCTs was divided by this prevalence.

Mapping

Maps were created in MapInfo (Version 8.0, MapInfo Ltd.) geographical information software using PCT primary care localities available from the UK Ordnance Survey Executive Agency.⁵

RESULTS

MSM population size 16–44 years in London

Based on NATSAL II estimates, 98,330 [75,088; 126,935] (1,787,818 × 5.5% [4.2–7.1%]) MSM aged 16–44 were living in Greater London in 2008. Of those 66,064 (50,806; 85,396), (798,839 × 8.27% [6.36–10.69%]) MSM aged 16–44 years were living in inner London compared with 36,790 [22,054; 60,723] (988,979 × 3.72% [2.23–6.14%] in outer London.

SOPHID distribution of MSM with diagnosed HIV infection across Greater London (all ages)

In 2008, 12,312 MSM with diagnosed HIV infection were living in Greater London. Of these, 9666 (79%) were living in an inner London PCT in 2008. Lambeth had the highest number of HIV-infected MSM (1655), followed by Southwark (1048), Westminster (904), Camden (885), Islington (823) and Kensington and Chelsea (764). Sixty-four percent (6250/9666) of all diagnosed MSM living in inner London were aged between 16 and 44 years (PCT range 54–72%). The numbers of MSM with diagnosed HIV were substantially lower in the outer London PCTs, with highest numbers in Waltham Forest (187), Ealing (161) and Greenwich (157). Fifty-nine percent (1581/2646) of MSM in outer London were aged 16 to 44 years. The analyses below are based on 6250 MSM with diagnosed HIV living in inner London who were aged between 16–44 years.

Estimated overall diagnosed HIV prevalence in MSM in inner London

The estimated overall prevalence of diagnosed HIV infection among MSM aged 16 to 44 years in inner London, calculated as above, was 9.5% (6,250/66,064; [95% CI: 7.3–12.3%]).

MSM population size estimates at PCT level

Table 1 shows the estimated number of MSM aged 16–44 years for each inner London PCT using the NoBE and SWE . The data show that while the overall estimate for inner London remains the same (subject to rounding), there is substantial variation at PCT level distribution between the two methods. The SWE method provides higher estimates of the resident MSM populations than the NoBE method for Lambeth (+5790 additional MSM), Southwark (+2028), Camden (+1519) and Islington (+1352) PCTs, but substantially lower estimates for Newham (−3495), Wandsworth (−3262), Haringey (−2401) and Lewisham (−2256). Differences between both estimates were statistically significant for Lambeth, Lewisham, Haringey, Newham and Wandsworth PCTs.

Table 1

Male population aged 16–44 years, number of MSM with diagnosed HIV aged 16–44 years, PCT-level population estimates of MSM aged 16–44 years according to the NoBE and SWE methods, and their differences according to inner London PCT of residence, 2008

London sector and PCT	Male population 16–44 years*	Number of MSM with diagnosed HIV 16–44 years^†	Estimated MSM 16–44 years		Difference (SWE − NoBE)	SWE % of MSM of male population
London sector and PCT	Male population 16–44 years*	Number of MSM with diagnosed HIV 16–44 years^†	NoBE^‡ (95% CI)	SWE^§ (95% CI)	Difference (SWE − NoBE)	SWE % of MSM of male population
SE Lambeth	79,289	1168	6557 (5043–8476)	12,347 (9496–16,000)	5790**	15.6
SE Southwark	71,344	750	5900 (4537–7627)	7928 (6098–10,274)	2028	11.1
SE Lewisham	67,672	316	5596 (4304–7234)	3340 (2569–4329)	(−) 2256**	4.9
NEC Camden	52,697	556	4358 (3352–5633)	5877 (4520–7616)	1519	11.2
NEC Islington	53,316	545	4409 (3391–5699)	5761 (4431–7466)	1352	10.8
NEC Tower Hamlets	64,296	466	5317 (4089–6873)	4926 (3789–6384)	(−) 391	7.7
NEC City & Hackney	57,861	414	4785 (3680–6185)	4376 (3366–5671)	(−) 409	7.6
NEC Haringey	59,962	242	4959 (3814–6410)	2558 (1967–3315)	(−) 2401**	4.3
NEC Newham	64,500	174	5334 (4102–6895)	1839 (1415–2384)	(−) 3495**	2.9
NW Westminster	57,130	532	4725 (3633–6107)	5624 (4325–7288)	899	9.8
NW Kensington & Chelsea	40,817	414	3376 (2596–4363)	4376 (3366–5671)	1001	10.7
NW Hammersmith & Fulham	48,847	347	4040 (3107–5222)	3668 (2821–4753)	(−) 372	7.5
SW Wandsworth	81,108	326	6708 (5158–8670)	3446 (2650–4466)	(−) 3262**	4.2
Inner London	798,839	6250	66,064 (50,806–85,396)	66,068 (50,813–85,616)	4	8.3

MSM = men who have sex with men; PCT = primary care trust; NoBE = NATSAL II only Based Estimates; SWE = SOPHID Weighted Estimates

*GLA – RLP High 2008 data

^†SOPHID data 2008

^‡NoBE: Proportion of MSM in inner London from NATSAL (8.27% [95% CI: 6.36–10.69%]) × estimated male population in each PCT (column 1)

^§SWE: Number of MSM with diagnosed HIV in each PCT (2)/Estimated overall diagnosed HIV prevalence in MSM in inner London (9.5% [95% CI: 7.3–12.3%])

**Statistically significant at 95% level

According to the SWE method, MSM accounted for an estimated 10–16% of the resident male population in Lambeth, Camden, Kensington and Chelsea, Islington, Southwark and Westminster, but substantially less than the inner London average (8.3%) in the other PCTs, particularly Newham Wandsworth, Lewisham and Haringey (2.9–4.9%) (Figure 1).

Figure 1

Estimated proportion of men who have sex with men among resident male population aged 16–44 years in inner London primary care trusts, 2008 (1) City and Hackney primary care trust (PCT); (2) Haringey PCT; (3) Islington PCT; (4) Camden PCT; (5) Westminster PCT; (6) Kensington and Chelsea PCT; (7) Hammersmith and Fulham PCT; (8) Wandsworth PCT; (9) Lambeth PCT; (10) Southwark PCT; (11) Lewisham PCT; (12) Tower Hamlets PCT; (13) Newham PCT

The map shows the geographical distribution of MSM populations and indicates that the central London PCTs Lambeth, Camden, Southwark, Islington and Kensington and Chelsea have the highest estimated proportions of MSM among the resident male population.

DISCUSSION

This study provides a novel method for estimating at-risk MSM population sizes in high HIV prevalence areas using the local distribution of MSM with diagnosed HIV infection for weighted estimates at the PCT level. Our SWEs provide greater heterogeneity in the number of MSM aged 16 to 44 years residing in inner London PCTs (range 1839 to 12,347) compared with NoBE (range 3376 to 6708) with four inner London PCTs having significantly different MSM population estimates. PCTs with estimated proportions of MSM above 10% were Lambeth, Southwark, Camden, Islington and Kensington and Chelsea.

There are several limitations to our study. Our estimated diagnosed HIV prevalence and MSM population size in inner London is based on denominators produced by a subset of the Greater London NATSAL II sample which is limited by small sample size (further limitations of NATSAL II discussed below) and GLA population estimates which are complex statistical population projections based on the 2001 Census. However, despite these limitations of using survey and population data our overall inner London HIV prevalence estimates are nevertheless consistent with findings of diagnosed HIV prevalence in recent London-based community surveys.^6,7 Neither the SWE or NoBE population estimates of MSM in inner London PCTs are likely to be completely accurate as these rely on oversimplified assumptions of constant diagnosed HIV prevalence among MSM or a constant proportion of MSM among the male population across all inner London PCTs, but we believe the assumptions are sufficiently reasonable for planning purposes. The assumption underlying the SWE was informed by the observation that the proportional breakdown of number of diagnosed HIV-infected MSM by PCT in SOPHID largely mirrors residential patterns of participants in the London sample of the National Gay Men's Sex Surveys⁸ and Gay Men's London Gym Surveys,⁹ (unpublished data, personal communication, Professor J Elford). Given universal open access to numerous sexual health and HIV services and excellent transport systems within inner London, it seems unlikely that the diagnosed HIV prevalence should vary greatly among MSM at a PCT level. Furthermore, our estimates based on diagnosed MSM assume that the ratio of diagnosed to undiagnosed cases is constant across PCTs, which may not be the case, although we have no evidence to suggest that it varies widely across different parts of London. In the absence of reference studies on small area geographical HIV prevalence and sexual mixing patterns these assumptions would need to be validated through further research. The 95% CIs for NoBE and SWE are presented to indicate the extent of the difference between the two estimates, but will need to be interpreted with caution due to the crudeness of source data discussed above.

SOPHID data are based on diagnosed persons accessing HIV care within a given year and some individuals may not access care every year; however, these numbers are very low for MSM.

NATSAL II has been the key to informing sexual health interventions in Britain; however, information is restricted to people aged 16–44 years (reflecting the female reproductive ages). Therefore, MSM population size estimates in this paper substantially underestimate the total MSM population. This may be a particular drawback, not only as more than one-third of MSM living with diagnosed HIV are now aged over 44 years and older MSM continue to be at risk of new infection, but also for the recognition of the wider health needs of the ageing cohort of MSM. Lastly, behavioural information in NATSAL is self-declared; comparison of data from NATSAL 1990 and 2000 demonstrated significant increases in the reported prevalence of several sexual behaviours among those aged 16–44 years.¹⁰ In line with further societal liberalization over the last decade and improved information collection techniques NATSAL 2010, which will also survey people aged 16–75 years, may be able to provide even more accurate estimates of the prevalence of same-sex behaviour.

However, despite these current limitations we believe our findings to be sufficiently reliable for health planning and commissioning purposes. Our findings are particularly relevant for Lambeth and Southwark, with SWE of more than 20,000 MSM aged 16–44 years – these are more likely to reflect the real number of MSM at risk in these PCTs compared with NoBE (12,500). These neighbouring PCTs, and in particular, the northern part of Lambeth PCT (Vauxhall), experienced significant growth in gay commercial venues in recent years and currently constitutes the hub of London's gay nightclub life: ‘Gay Vauxhall Village’¹¹ (estimated venue capacity 5000 people [personal communication, Lambeth licensing]). This conglomeration also facilitates social and sexual networks of MSM across London, the UK and Europe. We believe that large proportions of MSM among resident male populations (up to 16% in Lambeth) make the case for the commissioning of health services and programmes that address the sexual and wider health needs of MSM. In addition to targeted sexual health interventions (HIV prevention, access to HIV/STI testing, post-exposure prophylaxis following sexual exposure [PEPSE]) aimed at MSM, local PCTs could ensure, for example, that their health inequalities agenda tackles homophobia and HIV-related stigma among health and other professionals through staff education and training.^12–14 PCTs may wish to include MSM health needs in the established ‘Staying Healthy’ and inequality work streams such as mental health, smoking cessation, drug and alcohol, as ill health rates in MSM, especially diagnosed HIV-infected MSM, are well above those in the general population.^15–17

CONCLUSION

We provide a novel method of estimating the number of at-risk MSM living in inner London PCTs. Our findings indicate that the proportion of MSM may vary widely between PCTs. MSM constitute significant proportions of the resident male population in several PCTs. Understanding the true size of the local at-risk populations is crucial for health-care planning, particularly to enable appropriate inclusion of their health needs in core PCT prevention and service programming. In the context of data paucity and growing HIV prevalence in MSM, validation of our findings is needed, as well as an evaluation of the usefulness of population-based behavioural studies and HIV surveillance systems in respect of service planning.

References

Johnson

, Mercer

, Erens

, Sexual behaviour in Britain: partnerships, practices, and HIV risk behaviours. Lancet 2001;358:1835–42

Dodds

, Johnson

, Parry

, Mercey

. A tale of three cities: persisting high HIV prevalence, risk behaviour and undiagnosed infection in community samples of men who have sex with men. Sex Transm Infect 2007;83:392–6

Wilkinson

, Dodds

, Copas

, Sexual Health Survey of Gay Men. London: UCL, 2008

Elford

, Bolding

, Davis

, Trends in sexual behaviour among London homosexual men 1998–2003: implications for HIV prevention and sexual health promotion. Sex Transm Infect 2004;80:451–4

10.

Copas

, Wellings

, Erens

, The accuracy of reported sensitive sexual behaviour in Britain: exploring the extent of change 1990–2000. Sex Transm Infect 2002;78:26–30

11.

12.

13.

Branson

, Handsfield

, Lampe

, Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings. MMWR Recomm Rep 2008;55(RR-14):1–17

14.

Martin

, Roland

, Neilands

, Use of postexposure prophylaxis against HIV infection following sexual exposure does not lead to increases in high-risk behavior. AIDS 2004;18:787–92

15.

Harding

, Bensley

, Corrigan

. Targeting smoking cessation to high prevalence communities: outcomes from a pilot intervention for gay men. BMC Public Health 2004;4:43

16.

King

, Semlyen

, Tai

, A systematic review of mental disorder, suicide, and deliberate self harm in lesbian, gay and bisexual people. BMC Psychiatry 2008;8:70

17.

Wasted opportunities: problematic alcohol and drug use among gay men and bisexual men. Sigma Research , 2009. See http://www.sigmaresearch.org.uk (last accessed 18 December 2010)